Real-time positioning and order receiving system for home life service based on dynamic weight

Abstract

The application discloses a kind of real-time positioning and order receiving system of family life service based on dynamic weight, comprising: demand analysis and order generation module, to receive, parse and process the family life service demand of user, and generate structured service order, provide standardized demand data input for subsequent matching;Positioning and tracking module, to real-time acquisition, processing and analysis service personnel's geographic location information;Dynamic weight calculation module, to real-time calculation order matching weight and service personnel's order receiving weight, realize dynamic, multidimensional, real-time matching decision;Intelligent matching and order receiving module, based on dynamic weight calculation result realizes the intelligent matching of order and service personnel, and provides flexible order receiving mechanism;Two-way interaction and feedback module, to realize the two-way interaction between user and service personnel.The application significantly improves the accuracy and adaptability of matching.

Description

Technical Field

[0001] This invention relates to the field of information technology for home life services, and in particular to a real-time location and order-taking system for home life services based on dynamic weights. Background Technology

[0002] With the acceleration of urbanization and the increasing pace of life, household services are gradually transforming from traditional referral and telephone booking models to platform-based and digital models, with the supply and demand of services on online platforms continuing to grow. Household services now cover a variety of scenarios, including cleaning, appliance repair, housekeeping, in-home cooking, nannies / maternity nurses, and elderly care, and users' demands for the immediacy, accuracy, and personalization of these services are becoming increasingly urgent.

[0003] However, existing domestic service platforms still have the following significant shortcomings in terms of supply and demand matching efficiency and accuracy: The matching dimensions are too simplistic and lack dynamic weight adjustment mechanisms. Currently, most platforms rely on static service personnel tags and simple user descriptions entered during registration. When a user submits a request, a rule engine extracts keywords from the request and matches them with service personnel tags, recommending the matched service personnel to the user. This approach cannot dynamically adjust the weights of matching factors based on real-time changing factors (such as peak-hour congestion, sudden surges in demand, and the real-time busyness of service personnel), leading to significant discrepancies between the matching results and actual expectations, often requiring multiple rounds of communication to finalize the match. While some solutions introduce the concept of dynamic weights, these primarily target lead allocation scenarios, calculating weights based on real-time task status and historical performance data, without deeply integrating them with the specific scenarios of home life services (such as LBS real-time location, multi-dimensional matching of service skills, and real-time user feedback). Other solutions use a pairwise scoring calculation based on preset scheduling rules, assigning weights to each rule; however, these weights are mostly static preset values, lacking real-time adjustment capabilities.

[0004] The utilization rate of real-time location information is low, lacking deep integration across spatiotemporal dimensions. While existing platforms have partially incorporated LBS positioning technology to obtain user and service personnel location information, this is mostly used only for simple nearest-location order dispatch or route display, without deep coupling of location data with dynamic weighting algorithms. Supply and demand profiles are static, lacking real-time update mechanisms. Service personnel capability profiles remain at the registration stage, lacking updates based on dynamic data such as historical performance records, task types, and service ratings, resulting in information asymmetry between supply and demand.

[0005] The order-taking model is rigid and lacks a two-way interactive optimization mechanism. Existing systems mostly adopt a single manual order-grabbing or automatic order-dispatch mode, lacking a real-time two-way matching and interaction mechanism between users and service personnel. In the manual order-grabbing mode, order pushes lack priority sorting, requiring service personnel to manually sift through a large number of orders, resulting in low efficiency. In the automatic order-dispatch mode, the system is mostly a one-way decision-making process, with users unable to participate in the selection of order recipients. This rigid model fails to meet the requirements of familiarity, trust, and personalized preferences for service personnel in home services. Summary of the Invention

[0006] The purpose of this invention is to provide a real-time location and order-taking system for home life services based on dynamic weights, so as to solve the technical problem of how to achieve efficient, accurate, dynamic matching and order taking for home life services.

[0007] This invention is implemented using the following technical solution: a real-time location and order-taking system for home life services based on dynamic weights, comprising: The demand parsing and order generation module is used to receive, parse, and process users' household life service demands, and generate structured service orders, providing standardized demand data input for subsequent matching; The location and tracking module is used to acquire, process, and analyze the geographical location information of service personnel in real time; The dynamic weight calculation module is used to calculate the matching weight of an order and the order acceptance weight of a service personnel in real time based on service order and geographical location information, so as to realize dynamic, multi-dimensional and real-time matching decisions. The intelligent matching and order acceptance module uses dynamic weight calculation results to intelligently match orders with service personnel and provides a flexible order acceptance mechanism. The two-way interaction and feedback module is used to enable two-way interaction between users and service personnel, and to update the profiles of both supply and demand sides in real time based on the interaction data.

[0008] Furthermore, it also includes a monitoring and anomaly handling module, which is used to continuously monitor the system operation, detect and handle various abnormal situations, and periodically optimize the weight parameters through the collection of operating data and offline training to form a data closed loop.

[0009] Furthermore, the demand parsing and order generation module includes: a multimodal demand acquisition unit, which includes: The natural language parsing subunit is used to parse the service requirements described by users in natural language into structured service requirement elements through a large language model; The form-based requirement input sub-unit provides a standardized service requirement form interface for users to select or fill in service type, service address, appointment time and special requirements; The historical demand association sub-unit is used to predict users' potential service needs based on their historical order data, and generate service demand forecasts in advance accordingly.

[0010] Furthermore, the demand parsing and order generation module also includes: an order generation and storage unit, which includes: The order standardization subunit is used to standardize the demand elements output by the multimodal demand collection unit according to the preset order template, and generate a standardized work order that includes a unique order identifier, user ID, service type code, service address and geographic coordinates, appointment time window, service duration estimate, special requirements description, order priority level and order status. The order priority initialization subunit is used to calculate the initial priority level of an order based on the urgency of the order's appointment time, user level, and special needs. This initial priority level is adjusted in real time by the dynamic weight calculation module during the subsequent matching process. The order storage subunit uses a distributed database architecture to store standardized order data, supports high-concurrency read and write operations, and pushes changes to relevant modules in real time via a message queue.

[0011] Furthermore, the positioning and tracking module includes: The multi-source positioning data acquisition unit is used to collect the location information of service personnel in real time through a fusion of GPS / BeiDou positioning, base station assisted positioning, and Wi-Fi / Bluetooth assisted positioning. The location data processing and filtering unit is used to perform anomaly filtering, Kalman filtering smoothing, and map matching correction on the raw location data. The trajectory tracking and ETA prediction unit is used to generate continuous movement trajectories of service personnel, dynamically predict the estimated time of arrival at the service address, and monitor the status of service personnel entering and leaving the service area through geofencing.

[0012] Furthermore, the dynamic weight calculation module calculates the dynamic weight of the order based on spatial, temporal, user, and urgency dimensions, and calculates the dynamic weight of the service personnel based on skill matching, reputation score, load score, location advantage, and willingness to accept orders. Specifically, this includes: The multi-dimensional weighting factor acquisition unit is used to collect spatial dimension factors, time dimension factors, capability dimension factors, user preference dimension factors, and real-time dynamic factors in real time. The dynamic weight calculation engine is used to calculate the dynamic weight of the order side and the dynamic weight of the service personnel side according to the multi-dimensional weight factors, and to calculate the comprehensive matching score between the order and the service personnel based on the two. When calculating the comprehensive matching score, the dynamic weight calculation engine calculates the coupling coefficient according to the matching degree between the special requirements filled in the order and the actual qualifications of the service personnel, the similarity between the user's historical service preferences and the current service personnel profile, the order appointment time window and the service personnel's available time margin, and the service personnel's familiarity with the area where the service address is located. The comprehensive matching score is obtained by multiplying the product of the dynamic weight of the order side and the dynamic weight of the service personnel side by the coupling coefficient. The weight update and iterative optimization unit is used to dynamically adjust the weight parameters based on real-time feedback and system operation data.

[0013] Furthermore, the intelligent matching and order acceptance module includes: The multi-mode matching and scheduling unit supports global optimal matching mode, real-time greedy matching mode and hierarchical priority matching mode, and divides orders into different levels for differentiated scheduling based on order priority level; The two-way order-receiving engine supports both system-assigned order mode and service personnel-bid order mode. It dynamically selects between the assignment and bidding modes based on order attributes and the real-time status of service personnel. Urgent orders and VIP user orders are given priority to the system-assigned order mode, while routine orders are given priority to the bidding mode.

[0014] Furthermore, the service personnel order-grabbing mode in the two-way order-accepting engine is implemented as follows: After the orders are sorted according to the results of dynamic weight calculation, they are pushed to the terminals of qualified service personnel. First, candidate service personnel within a preset range of the order service address are selected based on the service personnel's location advantage score. Then, the candidate service personnel are sorted according to the comprehensive order-accepting weight of the service personnel, and the service personnel with the highest weight gets priority to grab the order. The order grabbing adopts a timed order grabbing mechanism. If the service personnel do not confirm the order within the preset response time, the next service personnel will be selected in order of weight.

[0015] Furthermore, the two-way interaction and feedback module includes: The service process visualization unit is used to display the location, travel route and estimated arrival time of service personnel to users in real time during the service process, and to synchronize the service progress with the service personnel. The evaluation feedback unit is used to collect multi-dimensional evaluations of service personnel from users and evaluations of users from service personnel after the service is completed, and to trigger the recalculation of service personnel reputation scores in the dynamic weight calculation module in real time by using the evaluation data. The real-time profile update unit is used to update user profiles and service personnel profiles in real time based on real-time demand text, historical order data, and evaluation behavior data through incremental learning algorithms. The updated profiles are immediately included in the calculation of dynamic weights.

[0016] Furthermore, the real-time profile update unit includes a lifecycle stage perception subunit; the lifecycle stage perception subunit is used to identify the user's family lifecycle stage based on the historical change trend of the user profile, and automatically adjust the recommendation strategy when a change in the user's lifecycle stage is detected, and actively push family life service types that match the current stage.

[0017] The beneficial effects of this invention are as follows: The dynamic weighting coefficients of this invention change in real time with time-based factors, regional factors, user factors, and real-time data feedback. They no longer rely on fixed rule configurations but can automatically adjust the weight allocation of each dimension according to the real-time scenario. This allows for priority to be given to time-sensitive orders during peak hours, priority to spatial matching in service-sparse areas, and differentiated services in high-level user scenarios, significantly improving the accuracy and adaptability of matching.

[0018] This invention utilizes technologies such as user lifecycle stage perception, real-time evaluation feedback, and incremental learning to achieve real-time updates of user profiles and service personnel profiles. The updated profiles are immediately included in the calculation of dynamic weights, which enables the matching accuracy to continuously improve as the system runs, effectively solving the problem of information asymmetry between supply and demand.

[0019] This invention supports three modes: global optimal matching, real-time greedy matching, and hierarchical priority matching, which can be flexibly switched according to order attributes and platform policies. The two-way order-taking engine supports the integration of system order dispatch and service personnel order grabbing. Urgent orders and VIP user orders are given priority by system dispatch to ensure response speed, while daily orders use the order grabbing mode to mobilize the initiative of service personnel. The order load balancing mechanism effectively prevents excessive load on service personnel and ensures service quality. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0021] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0023] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0024] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0025] See Figure 1 The real-time location and order-taking system for home life services based on dynamic weights includes: a demand analysis and order generation module, a location and tracking module, a dynamic weight calculation module, an intelligent matching and order-taking module, a two-way interaction and feedback module, and a monitoring and anomaly handling module.

[0026] In this embodiment, the demand parsing and order generation module is responsible for receiving, parsing, and processing users' household service demands, generating structured service orders, and providing standardized demand data input for subsequent matching. Specifically, it includes a multimodal demand acquisition unit and an order generation and storage unit.

[0027] The multimodal demand acquisition unit includes: The Natural Language Parsing Subunit employs multimodal demand parsing technology based on a large language model to acquire service demand information submitted by users on their terminal devices in natural language, and parses the natural language demands into structured service demand elements. Specifically, this subunit accurately parses the explicit and implicit demands of user orders using a fine-tuned multimodal large language model, outputting a structured demand feature matrix. For fuzzy demand information, this subunit predicts candidate demands with different user preferences based on initial features, and guides users to select target demands from these candidate demands by generating guiding information, thereby achieving gradual refinement of demands through natural language interaction.

[0028] The form-based service request input sub-unit provides a standardized service request form interface, allowing users to quickly submit service requests by selecting the service type (such as cleaning, appliance repair, housekeeping, in-home cooking, nanny / maternity nurse, etc.), filling in the service address, selecting the appointment time, and filling in special requirements. This serves as a supplement to natural language parsing and a means of verification.

[0029] Historical Demand Association Subunit: Based on users' historical order data, this subunit uses big data analytics to process and concatenate family structure information and historical order information to obtain a comprehensive feature vector. This comprehensive feature vector is then input into a multilayer perceptron to generate a lifecycle stage vector. Preference and role vectors are constructed for each family member. Using the lifecycle stage vector as a condition, the attention weights for each member's perception at each stage are calculated to predict the user's potential service needs. This subunit can identify the user's current family lifecycle stage (e.g., newlywed period, childcare period, school age, empty nest period, elderly care period, etc.) and generate service demand predictions in advance, providing a basis for proactive recommendations.

[0030] The order generation and storage unit includes: Order Standardization Subunit: Standardizes the demand elements output by the multimodal demand collection unit according to the preset order template to generate a standardized order containing the following fields: unique order identifier, user ID, service type code, service address and geographic coordinates, appointment time window, estimated service duration, description of special requirements, order priority level, order creation timestamp, and order status (pending matching / matched / in progress / completed / cancelled / abnormal).

[0031] Order priority initialization sub-unit: Calculates the initial priority based on the order attributes.

[0032] Order storage subunit: Employs a distributed database architecture to store standardized order data, supporting high-concurrency read / write and real-time querying. When an order status changes, it is pushed to the relevant modules in real time via a message queue.

[0033] The positioning and tracking module is responsible for acquiring, processing, and analyzing the geographical location information of service personnel in real time, providing spatiotemporal data support for dynamic weight calculation and intelligent matching. Specifically, it includes: a multi-source positioning data acquisition unit, a location data processing and filtering unit, and a trajectory tracking and ETA prediction unit.

[0034] The multi-source positioning data acquisition unit includes: GPS / BeiDou Positioning Subunit: This subunit uses the GPS or BeiDou positioning module built into the service personnel's terminal equipment to collect positioning data such as latitude and longitude coordinates, altitude, speed, direction angle, and horizontal accuracy factor from the service personnel at a preset frequency (default 5 seconds / time). Employing high-precision positioning technology from the 5G SA network, it achieves sub-meter positioning accuracy in outdoor scenarios.

[0035] Base station assisted positioning subunit: In indoor environments or urban canyon areas with weak GPS signals, the location information of service personnel is obtained through base station triangulation technology of the mobile communication network. This serves as a supplement and degradation solution to GPS positioning, ensuring the continuity of positioning. Data collected includes cell ID, signal strength, round-trip time, etc.

[0036] Wi-Fi / Bluetooth Assisted Positioning Subunit: In indoor service scenarios, a technical solution combining low-resolution millimeter-wave radar and Wi-Fi fingerprint positioning is adopted. By collecting Wi-Fi access point information (BSSID, signal strength RSSI, channel frequency, etc.) and Bluetooth beacon information connected to the service personnel's terminal devices, and matching them with a preset indoor fingerprint map, high-precision indoor positioning is achieved, with positioning errors controlled within 3 meters.

[0037] The location data processing and filtering unit includes: The abnormal data filtering subunit employs an anomaly detection algorithm based on velocity and acceleration to identify and filter out positioning jump points that exceed reasonable limits. Specifically, it calculates the instantaneous velocity based on the time and distance between two consecutive positioning points. If the instantaneous velocity exceeds a preset maximum speed threshold (e.g., a speed limit of 120 km / h on urban roads), the positioning point is determined to be an abnormal jump point and replaced with a Kalman filter prediction value.

[0038] Kalman filter smoothing subunit: The continuous positioning sequence is smoothed using the Kalman filter algorithm to reduce the impact of positioning noise on position accuracy.

[0039] Map matching and correction subunit: This subunit matches the filtered location points with the electronic map road network data, and uses a Hidden Markov Model (HMM) to project the location points to the most probable road locations. This subunit uses the location point sequence as the observed state and candidate roads as the hidden state, and solves for the maximum probability path using the Viterbi algorithm to complete the map matching and correction.

[0040] The trajectory tracking and ETA prediction unit includes: Continuous trajectory generation subunit: This subunit integrates distance, radial velocity, and azimuth and pitch angles of target objects that meet preset existence conditions from positioning point data, and generates a continuous movement trajectory of service personnel in three-dimensional space through trajectory tracking processing. It supports historical trajectory playback and real-time trajectory display.

[0041] The ETA real-time prediction subunit, based on real-time location data, historical travel data, real-time traffic information, and the movement speed of service personnel, employs a time-series prediction algorithm based on a hybrid expert model to dynamically predict the estimated time for service personnel to reach the target service address. ETA prediction comprehensively considers the following factors: the ratio of straight-line distance to path distance, the average speed of the current road segment, the historical speed distribution over a given period, and the impact coefficient of traffic events. ETA prediction values ​​are dynamically adjusted according to changes in service personnel location and real-time updates to traffic conditions, and are pushed to user terminals in real time through the GIS system, eliminating waiting anxiety.

[0042] Geofencing monitoring subunit: A dynamic geofence (default radius 200 meters) is set up around the service address. An imminent arrival notification is triggered when service personnel enter the geofence area; an anomaly alarm is triggered when service personnel leave the service area. The geofence radius can be dynamically adjusted based on service type and user preferences. For example, the geofence radius for emergency repair services can be set to 100 meters to improve response accuracy, while the geofence radius for routine cleaning services can be set to 300 meters to provide more sufficient arrival buffer time.

[0043] The dynamic weight calculation module is responsible for calculating the matching weight of orders and the order acceptance weight of service personnel in real time, realizing dynamic, multi-dimensional, and real-time matching decisions. Specifically, it includes: a multi-dimensional weight factor collection unit, a dynamic weight calculation engine, and a weight update and iterative optimization unit.

[0044] The multi-dimensional weighting factor collection unit includes: The spatial dimension factor collection subunit collects the real-time path distance (non-linear distance) between the current location of the service personnel and the service address, calculates the actual driving distance using the map service provider's path planning API, the estimated arrival time based on the ETA prediction module, the service supply and demand density index of the service personnel's current area (calculated based on the ratio of the number of orders to be matched in the area to the number of service personnel), and the preference distribution of the service personnel's historical service coverage areas.

[0045] The time-dimensional factor collection subunit collects real-time traffic congestion index for the current time period (such as morning rush hour, noon, evening rush hour, night, etc.); online time and continuous service time of service personnel; time difference between order appointment time and current time, used to assess the urgency of the order; and average order response speed of service personnel in the current time period.

[0046] The capability dimension factor collection subunit collects static and dynamic capability data of service personnel, including: skill tag matching degree, calculated based on the degree of matching between service type requirements and service personnel skill tags; historical service rating, based on user evaluation ratings of historical orders, using time decay weight (recent evaluations have higher weight than distant evaluations); order completion rate, the proportion of orders actually completed after historical order acceptance to the total number of orders accepted; complaint rate and dispute rate, the number of complaints and disputes in historical orders; and service time completion rate, the deviation rate between actual service time and estimated service time.

[0047] The user preference dimension factor collection subunit collects personalized preference data from users, including: historical service personnel preferences (the sequence of service personnel IDs selected in the user's historical orders and their evaluation scores); service time preferences (the user's preferred service time periods, such as weekends, weekday daytime, evening, etc.); service personnel attribute preferences (the user's preference settings for service personnel's gender, age, years of service experience, etc.); and repurchase intention index (the probability prediction of the user placing repeated orders with the same service personnel).

[0048] The real-time dynamic factor collection subunit collects dynamic data that changes in real time: the real-time busy status of service personnel (whether they are currently performing service tasks, the remaining time of the task, the continuous working time, etc.); real-time evaluation feedback, users' real-time evaluation of the service in progress (such as instant ratings during the service process); and real-time order acceptance willingness, the order acceptance willingness level (active / neutral / busy / resting) set by service personnel through their terminal devices.

[0049] The dynamic weight calculation engine includes: The dynamic weight calculation subunit on the order side calculates the weight score of each order in real time across various dimensions. This subunit specifically includes the following functions: Spatial dimension scoring: Calculates the order's score based on the service supply and demand density index of the area where the order's service address is located, the historical service coverage popularity of that area, and the spatial relationship between the order's service address and the distribution of surrounding service personnel. Time dimension scoring: Calculates the order's score based on the urgency reflected by the time difference between the order's appointment time and the current time, the historical order acceptance success rate for the current time period, and the impact of the current time period's traffic congestion index on dispatch timeliness. User dimension scoring: Calculates the order's score based on user level, historical user satisfaction, user repurchase rate, and user loyalty to the platform's services. Urgency scoring: Calculates the order's urgency score based on whether the order carries an emergency marker, the urgency of the appointment time window, and whether the service type is an emergency service (such as emergency appliance repair). Dynamic weight coefficient adjustment function: Based on real-time changes in time-of-day, regional, and user factors, as well as historical matching feedback, the function automatically adjusts the weighting of the four dimensions in the overall order weight. This ensures that the weight of the time dimension is automatically increased during peak hours, the weight of the spatial dimension is automatically increased in areas with sparse service personnel, and the weight of the user dimension is automatically increased in high-level user scenarios. Comprehensive weight output function: The function weights and merges the scores of each dimension with their corresponding dynamic weight coefficients to output a comprehensive order weight score. This score is updated in real-time as any dimension factor or weight coefficient changes. The personnel-side dynamic weight calculation subunit calculates the order-accepting weight of each online service personnel in real time across various dimensions. This subunit specifically includes the following functions: Skill Matching Score: Based on the matching degree between the service personnel's skill tags and the service type of the current order to be matched, skill level certification status, and whether they possess special service qualifications (such as deep cleaning certification, appliance repair qualification), the service personnel's skill matching score is calculated. Reputation Score Calculation: Based on the service personnel's historical service ratings, order completion rate, positive review rate, complaint rate, and dispute rate, a time decay mechanism is used to make recent evaluations have a higher impact than long-term evaluations, comprehensively calculating the service personnel's reputation score. Load Score Calculation: Based on the number of tasks currently in progress, the remaining service time of each task, continuous working time, and rest intervals, the current load level of the service personnel is assessed. The lower the load, the higher the score, avoiding pushing new orders to overworked service personnel. Location Advantage Score: Based on the real-time path distance between the service personnel's current location and the order service address, estimated arrival time, and traffic conditions, the location advantage of the service personnel relative to the order is assessed. The closer the distance and the shorter the estimated arrival time, the higher the score. Order Acceptance Willingness Score: Calculates an order acceptance willingness score based on the service personnel's actively set order acceptance willingness level (including active, average, busy, resting, etc.) and real-time online status via their terminal devices. Dynamic Weight Coefficient Adjustment: Automatically adjusts the proportion of the above five dimensions' scores in the overall order acceptance weight based on the professional skill requirements of the service type scenario (e.g., increasing the weight of skill dimension in appliance repair scenarios), time period factors (e.g., increasing the weight of location and willingness dimensions during late nights), and platform operation strategy parameters. Comprehensive Order Acceptance Weight Output: Weights and merges the above dimension scores with their corresponding dynamic weight coefficients to output a comprehensive order acceptance weight score for the service personnel. This score updates in real-time as the service personnel's status changes (e.g., location changes, becoming idle after completing a service, receiving new feedback, etc.).

[0050] Matching Score Calculation Subunit: Based on the dynamic weights of the order and the service personnel, this subunit calculates the comprehensive matching score for each pair. Specifically, this subunit includes the following functions: Weight Product Calculation Function: Multiplies the comprehensive weight score from the order side with the comprehensive order-acceptance weight score from the service personnel side, using this as the basis for the matching score. Coupling Coefficient Calculation Function: Calculates the coupling coefficient between the order and the service personnel based on the following multi-dimensional factors: the degree of matching between the special requirements filled in the order (e.g., requiring a female service personnel, requiring pet care experience, etc.) and the actual qualifications and profile of the service personnel; the similarity between the user's historical service preferences (characteristics of service personnel previously selected by the user, repurchase tendency) and the current service personnel profile; the time overlap and available time margin between the order's appointment time window and the service personnel's existing task schedule; and the service personnel's historical service coverage frequency and familiarity with the area where the service address is located. Comprehensive Matching Score Calculation Function: Multiplies the weight product by the coupling coefficient to obtain the final comprehensive matching score between the order and the service personnel. Matching score matrix generation function: For each order to be matched and each candidate service provider, the above calculation is repeated to generate a matching score matrix with orders as rows and service providers as columns. This matrix is ​​used by the intelligent matching and order acceptance module for order dispatching decisions or order-grabbing sorting. Real-time update function: When the order-side weight of any order changes, or the personnel-side weight of any service provider changes, or any factor in the coupling coefficient changes, the affected matching score is automatically recalculated, and the corresponding elements in the matching score matrix are updated.

[0051] The weight update and iterative optimization unit includes: Real-time weight update sub-unit: Monitors real-time changes in factors across all dimensions. When the value of any factor changes, it immediately triggers a recalculation of the relevant weights.

[0052] Periodic weight optimization subunit: This subunit optimizes the parameters of the weight function offline at fixed intervals (e.g., daily or weekly). Based on historical matching data and order fulfillment results, it uses optimization algorithms such as gradient descent to adjust the parameters of the weight coefficients, maximizing the correlation between the matching score and target indicators such as order success rate and user satisfaction.

[0053] Anomaly Weight Correction Subunit: When anomalies in matching results are detected (such as orders with high matching scores being ultimately rejected, or highly-rated service personnel not receiving orders for an extended period), the anomaly correction process is initiated to temporarily adjust relevant weight coefficients and prevent the continued occurrence of anomalies. The correction amount is determined jointly through preset rules and real-time data analysis.

[0054] The intelligent matching and order acceptance module is responsible for intelligently matching orders with service personnel based on dynamic weight calculation results, and provides a flexible order acceptance mechanism, which includes: a multi-mode matching and scheduling unit, a two-way order acceptance engine, and a post-order processing unit.

[0055] The multi-pattern matching scheduling unit includes: Global optimal matching mode: Using the Hungarian algorithm or the KM algorithm, the global optimal matching scheme is solved based on the matching score matrix within at least a single batch or the entire city, so as to achieve the overall optimal matching between work orders and service personnel. It is especially suitable for scheduling scenarios of daily batch orders.

[0056] Real-time greedy matching mode: Suitable for immediate demand orders. When a user submits an order, the system immediately calculates the matching score between the order and all candidate service providers, and selects the service provider with the highest matching score that exceeds the preset matching threshold (default 0.7) for push or assignment. If the highest score is lower than the threshold, it enters a waiting queue, waiting for the next service provider's status to change before recalculating the matching.

[0057] Layered Priority Matching Mode: A distributed architecture is used to implement a layered traffic distribution scheduling strategy. Orders are divided into three priority levels: high, medium, and low. Low-priority orders are matched and allocated within the local service area to reduce cross-level transmission time; medium-priority orders are matched within the region, allowing allocation across adjacent regions; high-priority orders are globally scheduled, invoking the globally optimal matching algorithm to prioritize the matching success rate of high-priority orders.

[0058] The two-way order-taking engine includes: System Dispatch Mode Subunit: Based on the matching score calculation, the system automatically assigns orders to the service personnel with the highest matching score. The dispatch decision comprehensively considers the following constraints: service personnel's service capability constraints (skill matching, qualification certification); time window constraints (matching the appointment time with the service personnel's available time); distance constraints (ETA does not exceed the user's maximum acceptable waiting time, defaulting to 90 minutes); workload constraints (the service personnel's current number of tasks does not exceed the maximum number of concurrent tasks, defaulting to 3); and user preference constraints (user's preference restrictions for specific service personnel).

[0059] The service personnel order-grabbing mode sub-unit: Orders are sorted according to dynamic weight calculation results and then pushed to the terminals of eligible service personnel. Order push is based on the following priority rules: First, candidate service personnel within a certain range (default 5 kilometers) from the order service address are selected based on the service personnel's (location advantage score); second, the candidate service personnel are sorted according to the comprehensive order acceptance weight, and the service personnel with the highest weight get priority to grab the order; the order grabbing adopts a timed order grabbing mechanism. Service personnel must confirm the order within the preset response time (default 30 seconds) after receiving the push. If the timeout is exceeded, the next service personnel will be selected according to the weight order.

[0060] Hybrid order-taking mode sub-unit: The platform supports dynamically selecting the order dispatch or order-grabbing mode based on order attributes and service personnel status: Urgent orders and VIP user orders are given priority to the system dispatch mode to ensure service response speed; daily orders are given priority to the order-grabbing mode to mobilize the initiative of service personnel; the system can automatically switch the push strategy based on the real-time order acceptance willingness of service personnel: service personnel with high willingness will receive order-grabbing pushes first, while service personnel with low willingness will receive fewer pushes.

[0061] The post-order processing unit includes: Order Status Update Subunit: After a successful order acceptance, the order status is updated to "Matched", the ID of the service personnel who accepted the order, the order acceptance time, the order acceptance mode (dispatch / grab), and the order acceptance confirmation information (including basic information of the service personnel, estimated arrival time, etc.) is pushed to the user terminal.

[0062] Order Failure Handling Subunit: When a service provider refuses to accept an order or fails to respond within the timeout period, the service provider is temporarily removed from the current order's candidate queue (preset cooldown time: 5 minutes), and the order is pushed to the next candidate service provider according to the matching score ranking. If all candidate service providers refuse or time out, the emergency dispatch process of the system monitoring and exception handling module is triggered.

[0063] Order Load Balancing Subunit: Monitors the order load of service personnel. When the load score of a service personnel is lower than the preset threshold (default is 0.3), it suspends the push of new orders to that service personnel and resumes the push after the service personnel completes the current task, so as to avoid the service personnel being overloaded and causing a decline in service quality.

[0064] The two-way interaction and feedback module is responsible for enabling two-way interaction between users and service personnel, and for updating the profiles of both supply and demand sides in real time based on the interaction data. It includes a service process visualization unit, a profile real-time update unit, and an evaluation and feedback unit.

[0065] The service process visualization unit includes: Real-time location sharing sub-unit: During the service process, GIS positioning technology allows users to view the real-time location, route, and estimated arrival time of service personnel within their terminal APP. Service personnel can also view the user's service address and contact information.

[0066] Service progress tracking sub-unit: Service personnel record key service milestones (start of service, completion of each service stage, end of service, etc.) through the terminal APP during the service process, and the service progress is displayed synchronously on the user terminal, enhancing service transparency and user experience.

[0067] Real-time communication subunit: Provides instant messaging and voice call functions for users and service personnel, supporting real-time communication during the service process, such as users supplementing service requests and service personnel confirming service details.

[0068] The real-time image update unit includes: Multi-dimensional evaluation data collection sub-unit: After the service is completed, multi-dimensional evaluation data is collected from users and service personnel respectively: User evaluation of service personnel: service skills score (1-5 points), service attitude score, punctuality score, professionalism score, overall satisfaction score, and written comments; Service personnel evaluation of users: service environment score, cooperation score, reasonableness of special needs score, and willingness to provide service again, etc.

[0069] Real-time data entry of evaluation data: After the evaluation data is submitted, it is written to the database in real time and triggers the weight update event of the dynamic weight calculation module, which immediately affects the calculation of (reputation score) and realizes the real-time adjustment of the weight of service personnel.

[0070] The abnormal rating review sub-unit marks and reviews rating data that is abnormal (such as a 1-point rating or a 5-point rating). If a user's rating deviates significantly from the average rating level of service personnel, the system triggers a review process, where platform operators manually review the rating or determine its validity based on preset rules.

[0071] The real-time image update unit includes: The real-time user profile update sub-unit updates user profiles in real time based on multi-source data such as real-time user request text, historical order data, and evaluation behavior data, using an incremental learning algorithm. User profiles include: distribution of preferred service types, characteristics of service-preferred personnel (age / gender / years of experience, etc.), preferred service time periods, price sensitivity, and service evaluation tendencies.

[0072] The service personnel profile is updated in real time. Based on dynamic data such as service personnel's historical performance records, task type distribution, service rating sequence, order completion rate, and complaint records, the service personnel profile is updated in real time. Specifically, a capability profile of service personnel is constructed through multimodal information, including explicit skill information (service type, skill level, qualification certificate) and implicit trait information (service style, communication ability, user evaluation characteristics).

[0073] Lifecycle Stage Awareness Subunit: Based on the historical trends of user profiles, this subunit identifies changes in the user's family lifecycle stage. When a change in a user's lifecycle stage is detected (such as transitioning from the honeymoon phase to the parenting phase), the system automatically adjusts its recommendation strategy, proactively pushing service types that match the current stage (such as switching from routine cleaning to postpartum care services).

[0074] The monitoring and anomaly handling module is responsible for real-time monitoring of the system's operating status and handling of various anomalies, including: an operating status monitoring unit, an anomaly detection and handling unit, and a data closed-loop iteration unit.

[0075] The operation status monitoring unit includes: System performance monitoring subunit: Real-time monitoring of core KPIs such as order matching response time, matching success rate, average order acceptance response time, and user satisfaction. An alarm is triggered when any metric exceeds a preset threshold.

[0076] Service Personnel Status Monitoring Subunit: Monitors service personnel's online status, busy status, geographical location, service duration, and other status information in real time. When a service personnel is offline for more than a preset time (default 10 minutes), they are automatically marked as offline and order pushes are stopped.

[0077] Order Status Monitoring Subunit: Monitors the status flow of each order in real time, marks orders that have been in a pending matching state for a long time (default is more than 30 minutes without matching), and triggers the emergency dispatch process.

[0078] The anomaly detection and handling unit includes: Matching Anomaly Detection Subunit: Detects matching results based on anomaly detection rules. Triggering processing is triggered when the following anomalies occur: a large number of matching pairs below the threshold exist in the matching score matrix; high-priority orders fail to match multiple times consecutively; the load distribution of service personnel in the matching results is severely uneven.

[0079] Scheduling Anomaly Handling Subunit: When a scheduling anomaly is detected, the following actions are automatically performed: temporarily adjust the matching threshold to expand the range of candidate service personnel; trigger the manual scheduling intervention process for manual allocation by operations personnel; and push delay notification information to users and provide the option to cancel the order.

[0080] Service Anomaly Handling Subunit: Handles abnormal situations during the service process: Service personnel no-show: Automatically triggers the re-matching process and issues a compensation coupon to the user; Service dispute: Initiates the dispute handling process, freezes order payment, and connects to customer service for manual handling; Location loss: When the service personnel's terminal loses location signal for more than a preset time (default 15 minutes), a security alarm is triggered.

[0081] The data closed-loop iteration unit includes: Operational data acquisition subunit: continuously collects various types of data generated during system operation, including matching decision data, performance result data, user evaluation data, service personnel feedback data, etc., forming a closed-loop data flow.

[0082] Offline training subunit: Based on the collected runtime data, the dynamic weight calculation model is retrained offline periodically (e.g., weekly). The training objective is to minimize the deviation between the matching score and user satisfaction. Supervised learning methods are used, with historical order fulfillment results as labels, to optimize the parameters of the weight function.

[0083] Online Model Update Subunit: Supports online hot updates of models, deploying optimized weight parameters to the production environment without interrupting system services, and enabling continuous iterative evolution of the model.

[0084] For the foregoing embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.

[0085] The above embodiments describe the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the invention should be within the protection scope of the appended claims.

Claims

1. A real-time location and order-taking system for home services based on dynamic weights, characterized in that: include: The demand parsing and order generation module is used to receive, parse, and process users' household life service demands, and generate structured service orders, providing standardized demand data input for subsequent matching; The location and tracking module is used to acquire, process, and analyze the geographical location information of service personnel in real time; The dynamic weight calculation module is used to calculate the matching weight of an order and the order acceptance weight of a service personnel in real time based on service order and geographical location information, so as to realize dynamic, multi-dimensional and real-time matching decisions. The intelligent matching and order acceptance module uses dynamic weight calculation results to intelligently match orders with service personnel and provides a flexible order acceptance mechanism. The two-way interaction and feedback module is used to enable two-way interaction between users and service personnel, and to update the profiles of both supply and demand sides in real time based on the interaction data.

2. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, It also includes a monitoring and anomaly handling module, which is used to continuously monitor the system operation, detect and handle various abnormal situations, and periodically optimize the weight parameters through the collection of running data and offline training to form a data closed loop.

3. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The demand parsing and order generation module includes: a multimodal demand acquisition unit, which includes: The natural language parsing subunit is used to parse the service requirements described by users in natural language into structured service requirement elements through a large language model; The form-based requirement input sub-unit provides a standardized service requirement form interface for users to select or fill in service type, service address, appointment time and special requirements; The historical demand association sub-unit is used to predict users' potential service needs based on their historical order data, and generate service demand forecasts in advance accordingly.

4. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 3, characterized in that, The demand analysis and order generation module further includes: an order generation and storage unit, which includes: The order standardization subunit is used to standardize the demand elements output by the multimodal demand collection unit according to the preset order template, and generate a standardized work order that includes a unique order identifier, user ID, service type code, service address and geographic coordinates, appointment time window, service duration estimate, special requirements description, order priority level and order status. The order priority initialization subunit is used to calculate the initial priority level of an order based on the urgency of the order's appointment time, user level, and special needs. This initial priority level is adjusted in real time by the dynamic weight calculation module during the subsequent matching process. The order storage subunit uses a distributed database architecture to store standardized order data, supports high-concurrency read and write operations, and pushes changes to relevant modules in real time via a message queue.

5. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The positioning and tracking module includes: The multi-source positioning data acquisition unit is used to collect the location information of service personnel in real time through a fusion of GPS / BeiDou positioning, base station assisted positioning, and Wi-Fi / Bluetooth assisted positioning. The location data processing and filtering unit is used to perform anomaly filtering, Kalman filtering smoothing, and map matching correction on the raw location data. The trajectory tracking and ETA prediction unit is used to generate continuous movement trajectories of service personnel, dynamically predict the estimated time of arrival at the service address, and monitor the status of service personnel entering and leaving the service area through geofencing.

6. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The dynamic weight calculation module calculates the dynamic weight of the order side based on spatial, temporal, user, and urgency dimensions, and calculates the dynamic weight of the service personnel side based on skill matching, reputation score, load score, location advantage, and willingness to accept orders. Specifically, it includes: The multi-dimensional weighting factor acquisition unit is used to collect spatial dimension factors, time dimension factors, capability dimension factors, user preference dimension factors, and real-time dynamic factors in real time. The dynamic weight calculation engine is used to calculate the dynamic weight of the order side and the dynamic weight of the service personnel side according to the multi-dimensional weight factors, and to calculate the comprehensive matching score between the order and the service personnel based on the two. When calculating the comprehensive matching score, the dynamic weight calculation engine calculates the coupling coefficient according to the matching degree between the special requirements filled in the order and the actual qualifications of the service personnel, the similarity between the user's historical service preferences and the current service personnel profile, the order appointment time window and the service personnel's available time margin, and the service personnel's familiarity with the area where the service address is located. The comprehensive matching score is obtained by multiplying the product of the dynamic weight of the order side and the dynamic weight of the service personnel side by the coupling coefficient. The weight update and iterative optimization unit is used to dynamically adjust the weight parameters based on real-time feedback and system operation data.

7. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The intelligent matching and order acceptance module includes: The multi-mode matching and scheduling unit supports global optimal matching mode, real-time greedy matching mode and hierarchical priority matching mode, and divides orders into different levels for differentiated scheduling based on order priority level; The two-way order-receiving engine supports both system-assigned order mode and service personnel-bid order mode. It dynamically selects between the assignment and bidding modes based on order attributes and the real-time status of service personnel. Urgent orders and VIP user orders are given priority to the system-assigned order mode, while routine orders are given priority to the bidding mode.

8. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 7, characterized in that, The service personnel order-grabbing mode in the two-way order-receiving engine is implemented as follows: After the orders are sorted according to the results of dynamic weight calculation, they are pushed to the terminals of qualified service personnel. First, candidate service personnel within a preset range of the order service address are selected based on the service personnel's location advantage score. Then, the candidate service personnel are sorted according to the comprehensive order-receiving weight of the service personnel, and the service personnel with the highest weight gets priority to grab the order. The order grabbing adopts a timed order grabbing mechanism. If the service personnel do not confirm the order within the preset response time, the next service personnel will be selected in order of weight.

9. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The two-way interaction and feedback module includes: The service process visualization unit is used to display the location, travel route and estimated arrival time of service personnel to users in real time during the service process, and to synchronize the service progress with the service personnel. The evaluation feedback unit is used to collect multi-dimensional evaluations of service personnel from users and evaluations of users from service personnel after the service is completed, and to trigger the recalculation of service personnel reputation scores in the dynamic weight calculation module in real time by using the evaluation data. The real-time profile update unit is used to update user profiles and service personnel profiles in real time based on real-time demand text, historical order data, and evaluation behavior data through incremental learning algorithms. The updated profiles are immediately included in the calculation of dynamic weights.

10. The real-time location and order-taking system for home life services based on dynamic weights as described in claim 1, characterized in that, The real-time profile update unit includes a lifecycle stage perception subunit; the lifecycle stage perception subunit is used to identify the user's family lifecycle stage based on the historical change trend of the user profile, and automatically adjust the recommendation strategy when a change in the user's lifecycle stage is detected, and actively push family life service types that match the current stage.

Images